Signal transformation
Raw observations are transformed into bounded internal signal inputs using level, velocity, abnormality, and persistence features.
Methodology
CrisisMeter is an interpreted pressure engine, not a raw feed mirror.
Signals are validated, normalized, grouped into pressure domains, adjusted for breadth and contradiction, and translated into a public-facing pressure score.
Domain inference
Signals are combined inside weighted domains such as energy, financial conditions, logistics, and event pressure before regional modifiers are applied.
Stability controls
Elastic damping is applied against prior latent pressure so the public score stays stable while still responding to meaningful new pressure.
Public interpretation
The product publishes an interpreted score, state label, and explanatory context instead of exposing raw internal decomposition directly.